- Email: [email protected]
SS8.01: Diagnostic decision making in patients with suspected venous thromboembolism across hospital walls and age limits
S38
Symposia / European Geriatric Medicine 5S1 (2014) S17–S43
controlled trial to evaluate the effect of CGA in frail older patients identified amongst acute medical admissions – and the development of the Silver Book – a national guideline for the effective management of acutely unwell older patients. He will describe the journey from the research to implementation locally, regionally and nationally and how acute care for frail older patients is changing as a consequence of research findings. SS8.01 Diagnostic decision making in patients with suspected venous thromboembolism across hospital walls and age limits G.J. Geersing, H.L. Koek, H.J. Schouten, J.J.M. Van Delden UMC Utrecht, The Netherlands How diagnosing VTE in older patients might differ from diagnosing VTE in younger adult patients (G.J. Geersing): To correctly exclude the presence of VTE without need for further diagnostic work-up, so-called diagnostic decision rules – based on a weighed combination of signs and symptoms and the result of the D-dimer test – have been developed. These strategies have been derived and validated in both primary and secondary care patients suspected of VTE. Notably frail older patients might benefit from such a strategy provided that it can safely rule-out VTE in a substantial proportion of them without needing to be referred for imaging examination. Yet, the accuracy of these existing clinical decision rules to rule-out VTE has never been tested in elderly populations. Geert-Jan Geersing will discuss how diagnosing VTE in older patients might differ from diagnosing VTE in younger adult patients. The predictive performance of clinical decision rules is susceptible to changes in patient populations and these rules might therefore perform worse in older patients in whom the prevalence of both VTE and co-morbidity are higher and the presentation of VTE might be more obscure. Also, the translation of rules derived in hospital setting to primary care or nursing-home setting might be problematic. In addition, current available diagnostic strategies recommend referral for further imaging examination for more than half of the patients, whereas diagnostic decision strategies that would spare higher proportions of older patients the possible hazardous referral for imaging examination might better serve their needs. Based on: Schouten HJ, Koek HL, Moons KG, van Delden JJ, Oudega R, Geersing GJ. Eur J Gen Pract. 2013 Jun; 19(2): 123–7. Validity of clinical decision rules to rule out VTE in older ambulatory patients (H.L. Koek): Dineke Koek will present the results of the “Venous thromboembolism in the elderly” study; a prospective validation study on the accuracy of clinical decision rules to exclude venous thromboembolism in frail older nursing home patients and primary care patients (mean age 80 years) with clinically suspected deep vein thrombosis or pulmonary embolism. VTE occurred in 29% of the patients primarily suspected of pulmonary embolism and in 47% of those primarily suspected of deep vein thrombosis. This prevalence was much higher than in previous studies in populations of younger adult patients (reporting a prevalence between 7% and 20%). This resulted in a higher failure rate (false negative rate) in patients who had a low score on the clinical decision rule and a normal D-dimer test (6% in our study versus below 2% in previous studies). Dineke Koek will also discuss the potency of clinical decision rules to rule in VTE in frail older patients (as opposed to the current approach of ruling out VTE). A combined rule-out and rule-in approach may enable clinicians’ decision-making for up to 58% of patients without the need for further diagnostic work-up. Based on: Schouten HJ, Koek HL, Oudega R, Van Delden JJ, Moons KG, Geersing GJ. Accuracy of decision strategies in diagnosing deep vein thrombosis in frail older out-of-hospital patients – a validation study. Submitted. And: Schouten HJ, Geersing GJ, Oudega R, Van Delden JJ,
Moons KG, Koek HL. Accuracy of the Wells-rule for pulmonary embolism in older ambulatory patients. Submitted. The diagnostic value of the D-dimer test using either conventional or age-adjusted cut-off values in older patients with suspected VTE (H.J. Schouten): A normal D-dimer test can rule out VTE in patients with a nonhigh clinical probability according to a clinical decision rule. Since D-dimer levels increase with age, D-dimer testing is less useful to exclude VTE in older patients if the conventional cut-off value (500 mg/L) above which the test is considered abnormal is applied. As potential solution of this problem, researchers proposed to use an age-adjusted cut-off value (age·10 mg/L) in patients >50 years. In the third part of the symposium, Henrike Schouten will discuss the results of a systematic review and bivariate random effects metaanalysis on this topic. We included 13 cohorts that enrolled older patients suspected of VTE in whom D-dimer testing (using both conventional and age-adjusted cut-off values) and reference testing were performed. Based on published data we reconstructed 2x2 tables, stratified by predefined age-categories and applied D-dimer cut-off value. We found that the proportion of patients with a nonhigh clinical probability (according to a clinical decision rule) in whom D-dimer testing could exclude VTE was only 12.4% in those aged more than 80 years. Therefore, D-dimer testing has limited utility in older patients when the conventional cut-off value is applied. Application of age-adjusted cut-off values increased the specificity without modifying the sensitivity which remained >97% in all age categories and would have resulted in correctly avoided imaging examinations in 30–42% of patients over 60 years with a non-high probability as compared to 12–33% when the conventional cut-off value was applied. Based on: Schouten HJ, Koek HL, Oudega R, Geersing GJ, Janssen KJ, van Delden JJ, Moons KG. BMJ 2012 Jun 6; 344: e2985. And: Schouten HJ, Geersing GJ, Koek HL, Zuithoff NP, Janssen KJ, Douma RA, van Delden JJ, Moons KG, Reitsma JB. BMJ. 2013 May 3; 346: f2492. Considerations in decisions to either refer for- or to withhold additional diagnostic investigations in frail older patients (J.J.M. van Delden): Patients with a high risk of VTE require appropriate imaging examination to confirm or refute the diagnosis. These imaging modalities are mostly not available in primary care and nursing home settings, necessitating patients in the high-risk category to be referred to a hospital. Prior work has shown that frail older patients are vulnerable for distress and complications resulting from transitions to hospital-care. Hence, physicians might feel reluctant to refer frail elderly patients for additional investigations. Hans van Delden will set out the results of a study on physicians’ considerations in their decision-making to either refer for or to withhold additional diagnostic investigations in nursing home patients with suspected VTE. We applied both quantitative and qualitative methods. In the quantitative part, patient outcomes were related to the decision to withhold diagnostic investigations. Referral for additional diagnostic investigations was withheld in four out of ten nursing home patients for whom imaging examination for suspected VTE was indicated. Patients in whom diagnostic investigations were withheld had a higher mortality rate than referred patients. For a better understanding of elderly care physicians’ decisions, in-depth interviews were performed and analysed using the grounded theory approach. In their decisions to forgo diagnostic investigations, physicians incorporated the severity of symptoms and estimated prognosis of the disease in the light of the patients’ chronic condition, potential benefits of diagnostic investigations and whether perfor-ming investigations agreed with pre-established management goals in advance care planning. Based on: Schouten HJ, van Ginkel S, Koek HL, Geersing GJ, Oudega R, Moons KG, Van Delden JJ. J Am Med Dir Assoc. 2012 Oct; 13(8): 682–7.
Symposia / European Geriatric Medicine 5S1 (2014) S17–S43
And: Schouten HJ, Koek HL, Kruisman-Ebbers M, Geersing GJ, Oudega R, Kars MC, Moons KG, Van Delden JJ. Decisions to withhold diagnostic investigations in nursing home patients with a clinical suspicion of VTE. PlosOne. 2014. Accepted for publication. SS8.02 Implementing ageing research in primary and secondary care: The UK Ageing Research Network Model C. Rajkumar1 , J.T. Timeyin2 , K.M. Ali1 , S.T. Parker3 Brighton and Sussex Medical School, Brighton, United Kingdom; 2 Brighton and Sussex University Hospital Trust, Brighton, United Kingdom; 3 Sheffield Institute for studies on ageing, Sheffield, United Kingdom 1
Why do we need a Research Network? (C. Rajkumar): The talk will inform the audience of the benefits of a Comprehensive Local Research Network (CLRN) in Surrey and Sussex in relation to recruiting to large multi-centre studies with objective performance outcomes (recruiting to time and target). Professor Rajkumar will highlight the beneficial role of CLRN in working with the UK research governance mechanisms resulting in non-academic partners embracing the research agenda. Running a multicentre trial: The Probiotics experience (J.T. Timeyin): The Probiotic study (Probiotics for preventing antibiotic associated diarrhoea including closridium difficile infection) recruited 1127 patients from 28 hospitals across the UK. Ms Timeyin will present how interaction with the national CLRN has enhanced recruitment and ensured that the study sample size is achieved. Patient and Public Involvement (PPI) in supporting ageing research in the UK (K.M. Ali): The concept of engaging older people (patients and their carers) in all stages of clinical research is a unique UK priority. In the UK model, patient and public are equal partners in the research enterprise with a valuable contribution to research conduct, delivery, dissemination and implementation. Specific examples of successful PPI will be presented such as the Stroke Oxygen Study experience whereby meaningful PPI resulted in recruiting 8000 patients to the study. Overview of the UK Ageing Research Network (S.T. Parker): This talk will present the UK Ageing Research Network and how it facilitated successful recruitment to large multi-centre studies. By presenting the UK model and its links with the European strategic ageing research agenda a platform for future collaborations will be established. SS8.03 Hip fractures in Norway and the Netherlands: Can we learn something about the underlying risk factors? K. Holvik1 , N.M. van Schoor2 , M.S. Bakken1 , O.M. Steihaug1 1 University of Bergen, Bergen, Norway; 2 VU University Medical Center, Amsterdam, The Netherlands Hip fractures in Norway and The Netherlands (K. Holvik): Between-population comparisons can produce knowledge about factors affecting fracture risk that vary geographically, and provide indications as to whether differences in lifestyle, environment or ethnic and genetic constitution may explain the geographic variations. Such information is important for planning preventive measures on a public health level. We have identified incident hip fractures in older people participating in two community-based studies in Norway and the Netherlands: the Oslo Health Study (HUBRO), 59°N and the
S39
Longitudinal Aging Study Amsterdam (LASA), 52°N, respectively. These were large multipurpose studies collecting anthropometric measurements and lifestyle and background data from older individuals living at home. 25-hydroxyvitamin D and other blood sample markers were also collected in these studies. We will present descriptive data on hip fracture incidence in European countries, with focus on Norway and the Netherlands, and between-country differences and similarities in risk factors for hip fracture. We also plan to explore whether immigrants to Norway from the Netherlands and its neighboring countries maintain their lower fracture risk after migration, or whether these populations approach the fracture incidence of the high-risk host population. Longitudinal changes in vitamin D levels during aging in The Netherlands (N.M. van Schoor): Purpose: Longitudinal changes in serum 25-OHD levels during aging have not been studied extensively. When studying the longitudinal change, it is highly important to adequately adjust for seasonal variation. The objectives are: (1) To examine longitudinal changes in serum 25-OHD levels during aging; and (2) To describe the seasonal variation using a cosine function. Methods: Data of the Longitudinal Aging Study Amsterdam (LASA) were used, an ongoing cohort study. Two different cohorts were included: (1) younger cohort: aged 55–65 years at baseline, n = 738, follow-up of six years; (2) older cohort: aged 65–88 years at baseline, n = 1320, follow-up of thirteen years. Results: In the younger cohort, a longitudinal increase in mean serum 25-OHD levels of 4 nmol/L in six years was observed; in the older cohort, a longitudinal decrease of 4 nmol/L in thirteen years was observed. The seasonal variation was ±12 nmol/L in the younger cohort, and ±7 nmol/L in the older cohort. Conclusions: Long term serum 25-hydroxyvitamin D levels remained fairly stable during ageing with slightly increasing levels in persons aged 55–65 years, and slightly decreasing levels in persons aged 65–88 years. On average, the seasonal variation was larger than the longitudinal change. Associations between hip fractures and psychotropic drugs (M.S. Bakken): Psychotropic drugs are widely used and may cause injurious falls. Aim: To examine associations between the use of antidepressants, anxiolytics and hypnotics among older people and the risk of hip fracture. Methods: A nationwide prospective cohort study of people in Norway born before 1945 (n = 906,422) was conducted. We obtained information on all prescriptions of antidepressants, anxiolytics and hypnotics dispensed in 2004–2010 (the Norwegian Prescription Database) and all primary hip fractures in 2005–2010 (the Norwegian Hip Fracture Registry). We compared the incidence rates of hip fracture during drug exposure and non-exposure by calculating the standardized incidence ratio (SIR). Results: Altogether, 39,938 people (4.4%) experienced a hip fracture. The risk of hip fracture was increased for people exposed to any antidepressant, SIR 1.7 (95% confidence interval 1.7–1.8); tricyclic antidepressants, SIR 1.4 (1.3–1.5) and selective serotonin reuptake inhibitors (SSRIs), SIR 1.8 (1.7–1.8). Preliminary analysis revealed an increased risk of hip fracture among people exposed to anxiolytics, SIR 1.4 (1.4–1.5) and hypnotics, SIR 1.2 (1.1–1.2); the excess risk was highest regarding short-acting benzodiazepine (SAB) anxiolytics, SIR 1.5 (1.4–1.6). Benzodiazepine-like hypnotics (z-hypnotics) were associated with higher excess risk of hip fracture at night, SIR 1.3 (1.2–1.4) than during the day, SIR 1.1 (1.1–1.2). Conclusions: Our nationwide study shows an excess risk of hip fracture among older people using antidepressants, anxiolytics or hypnotics, including the widely prescribed SSRIs, SABs and z-hypnotics (particularly at night). Thus, cautious prescribing is needed.
Symposia / European Geriatric Medicine 5S1 (2014) S17–S43
controlled trial to evaluate the effect of CGA in frail older patients identified amongst acute medical admissions – and the development of the Silver Book – a national guideline for the effective management of acutely unwell older patients. He will describe the journey from the research to implementation locally, regionally and nationally and how acute care for frail older patients is changing as a consequence of research findings. SS8.01 Diagnostic decision making in patients with suspected venous thromboembolism across hospital walls and age limits G.J. Geersing, H.L. Koek, H.J. Schouten, J.J.M. Van Delden UMC Utrecht, The Netherlands How diagnosing VTE in older patients might differ from diagnosing VTE in younger adult patients (G.J. Geersing): To correctly exclude the presence of VTE without need for further diagnostic work-up, so-called diagnostic decision rules – based on a weighed combination of signs and symptoms and the result of the D-dimer test – have been developed. These strategies have been derived and validated in both primary and secondary care patients suspected of VTE. Notably frail older patients might benefit from such a strategy provided that it can safely rule-out VTE in a substantial proportion of them without needing to be referred for imaging examination. Yet, the accuracy of these existing clinical decision rules to rule-out VTE has never been tested in elderly populations. Geert-Jan Geersing will discuss how diagnosing VTE in older patients might differ from diagnosing VTE in younger adult patients. The predictive performance of clinical decision rules is susceptible to changes in patient populations and these rules might therefore perform worse in older patients in whom the prevalence of both VTE and co-morbidity are higher and the presentation of VTE might be more obscure. Also, the translation of rules derived in hospital setting to primary care or nursing-home setting might be problematic. In addition, current available diagnostic strategies recommend referral for further imaging examination for more than half of the patients, whereas diagnostic decision strategies that would spare higher proportions of older patients the possible hazardous referral for imaging examination might better serve their needs. Based on: Schouten HJ, Koek HL, Moons KG, van Delden JJ, Oudega R, Geersing GJ. Eur J Gen Pract. 2013 Jun; 19(2): 123–7. Validity of clinical decision rules to rule out VTE in older ambulatory patients (H.L. Koek): Dineke Koek will present the results of the “Venous thromboembolism in the elderly” study; a prospective validation study on the accuracy of clinical decision rules to exclude venous thromboembolism in frail older nursing home patients and primary care patients (mean age 80 years) with clinically suspected deep vein thrombosis or pulmonary embolism. VTE occurred in 29% of the patients primarily suspected of pulmonary embolism and in 47% of those primarily suspected of deep vein thrombosis. This prevalence was much higher than in previous studies in populations of younger adult patients (reporting a prevalence between 7% and 20%). This resulted in a higher failure rate (false negative rate) in patients who had a low score on the clinical decision rule and a normal D-dimer test (6% in our study versus below 2% in previous studies). Dineke Koek will also discuss the potency of clinical decision rules to rule in VTE in frail older patients (as opposed to the current approach of ruling out VTE). A combined rule-out and rule-in approach may enable clinicians’ decision-making for up to 58% of patients without the need for further diagnostic work-up. Based on: Schouten HJ, Koek HL, Oudega R, Van Delden JJ, Moons KG, Geersing GJ. Accuracy of decision strategies in diagnosing deep vein thrombosis in frail older out-of-hospital patients – a validation study. Submitted. And: Schouten HJ, Geersing GJ, Oudega R, Van Delden JJ,
Moons KG, Koek HL. Accuracy of the Wells-rule for pulmonary embolism in older ambulatory patients. Submitted. The diagnostic value of the D-dimer test using either conventional or age-adjusted cut-off values in older patients with suspected VTE (H.J. Schouten): A normal D-dimer test can rule out VTE in patients with a nonhigh clinical probability according to a clinical decision rule. Since D-dimer levels increase with age, D-dimer testing is less useful to exclude VTE in older patients if the conventional cut-off value (500 mg/L) above which the test is considered abnormal is applied. As potential solution of this problem, researchers proposed to use an age-adjusted cut-off value (age·10 mg/L) in patients >50 years. In the third part of the symposium, Henrike Schouten will discuss the results of a systematic review and bivariate random effects metaanalysis on this topic. We included 13 cohorts that enrolled older patients suspected of VTE in whom D-dimer testing (using both conventional and age-adjusted cut-off values) and reference testing were performed. Based on published data we reconstructed 2x2 tables, stratified by predefined age-categories and applied D-dimer cut-off value. We found that the proportion of patients with a nonhigh clinical probability (according to a clinical decision rule) in whom D-dimer testing could exclude VTE was only 12.4% in those aged more than 80 years. Therefore, D-dimer testing has limited utility in older patients when the conventional cut-off value is applied. Application of age-adjusted cut-off values increased the specificity without modifying the sensitivity which remained >97% in all age categories and would have resulted in correctly avoided imaging examinations in 30–42% of patients over 60 years with a non-high probability as compared to 12–33% when the conventional cut-off value was applied. Based on: Schouten HJ, Koek HL, Oudega R, Geersing GJ, Janssen KJ, van Delden JJ, Moons KG. BMJ 2012 Jun 6; 344: e2985. And: Schouten HJ, Geersing GJ, Koek HL, Zuithoff NP, Janssen KJ, Douma RA, van Delden JJ, Moons KG, Reitsma JB. BMJ. 2013 May 3; 346: f2492. Considerations in decisions to either refer for- or to withhold additional diagnostic investigations in frail older patients (J.J.M. van Delden): Patients with a high risk of VTE require appropriate imaging examination to confirm or refute the diagnosis. These imaging modalities are mostly not available in primary care and nursing home settings, necessitating patients in the high-risk category to be referred to a hospital. Prior work has shown that frail older patients are vulnerable for distress and complications resulting from transitions to hospital-care. Hence, physicians might feel reluctant to refer frail elderly patients for additional investigations. Hans van Delden will set out the results of a study on physicians’ considerations in their decision-making to either refer for or to withhold additional diagnostic investigations in nursing home patients with suspected VTE. We applied both quantitative and qualitative methods. In the quantitative part, patient outcomes were related to the decision to withhold diagnostic investigations. Referral for additional diagnostic investigations was withheld in four out of ten nursing home patients for whom imaging examination for suspected VTE was indicated. Patients in whom diagnostic investigations were withheld had a higher mortality rate than referred patients. For a better understanding of elderly care physicians’ decisions, in-depth interviews were performed and analysed using the grounded theory approach. In their decisions to forgo diagnostic investigations, physicians incorporated the severity of symptoms and estimated prognosis of the disease in the light of the patients’ chronic condition, potential benefits of diagnostic investigations and whether perfor-ming investigations agreed with pre-established management goals in advance care planning. Based on: Schouten HJ, van Ginkel S, Koek HL, Geersing GJ, Oudega R, Moons KG, Van Delden JJ. J Am Med Dir Assoc. 2012 Oct; 13(8): 682–7.
Symposia / European Geriatric Medicine 5S1 (2014) S17–S43
And: Schouten HJ, Koek HL, Kruisman-Ebbers M, Geersing GJ, Oudega R, Kars MC, Moons KG, Van Delden JJ. Decisions to withhold diagnostic investigations in nursing home patients with a clinical suspicion of VTE. PlosOne. 2014. Accepted for publication. SS8.02 Implementing ageing research in primary and secondary care: The UK Ageing Research Network Model C. Rajkumar1 , J.T. Timeyin2 , K.M. Ali1 , S.T. Parker3 Brighton and Sussex Medical School, Brighton, United Kingdom; 2 Brighton and Sussex University Hospital Trust, Brighton, United Kingdom; 3 Sheffield Institute for studies on ageing, Sheffield, United Kingdom 1
Why do we need a Research Network? (C. Rajkumar): The talk will inform the audience of the benefits of a Comprehensive Local Research Network (CLRN) in Surrey and Sussex in relation to recruiting to large multi-centre studies with objective performance outcomes (recruiting to time and target). Professor Rajkumar will highlight the beneficial role of CLRN in working with the UK research governance mechanisms resulting in non-academic partners embracing the research agenda. Running a multicentre trial: The Probiotics experience (J.T. Timeyin): The Probiotic study (Probiotics for preventing antibiotic associated diarrhoea including closridium difficile infection) recruited 1127 patients from 28 hospitals across the UK. Ms Timeyin will present how interaction with the national CLRN has enhanced recruitment and ensured that the study sample size is achieved. Patient and Public Involvement (PPI) in supporting ageing research in the UK (K.M. Ali): The concept of engaging older people (patients and their carers) in all stages of clinical research is a unique UK priority. In the UK model, patient and public are equal partners in the research enterprise with a valuable contribution to research conduct, delivery, dissemination and implementation. Specific examples of successful PPI will be presented such as the Stroke Oxygen Study experience whereby meaningful PPI resulted in recruiting 8000 patients to the study. Overview of the UK Ageing Research Network (S.T. Parker): This talk will present the UK Ageing Research Network and how it facilitated successful recruitment to large multi-centre studies. By presenting the UK model and its links with the European strategic ageing research agenda a platform for future collaborations will be established. SS8.03 Hip fractures in Norway and the Netherlands: Can we learn something about the underlying risk factors? K. Holvik1 , N.M. van Schoor2 , M.S. Bakken1 , O.M. Steihaug1 1 University of Bergen, Bergen, Norway; 2 VU University Medical Center, Amsterdam, The Netherlands Hip fractures in Norway and The Netherlands (K. Holvik): Between-population comparisons can produce knowledge about factors affecting fracture risk that vary geographically, and provide indications as to whether differences in lifestyle, environment or ethnic and genetic constitution may explain the geographic variations. Such information is important for planning preventive measures on a public health level. We have identified incident hip fractures in older people participating in two community-based studies in Norway and the Netherlands: the Oslo Health Study (HUBRO), 59°N and the
S39
Longitudinal Aging Study Amsterdam (LASA), 52°N, respectively. These were large multipurpose studies collecting anthropometric measurements and lifestyle and background data from older individuals living at home. 25-hydroxyvitamin D and other blood sample markers were also collected in these studies. We will present descriptive data on hip fracture incidence in European countries, with focus on Norway and the Netherlands, and between-country differences and similarities in risk factors for hip fracture. We also plan to explore whether immigrants to Norway from the Netherlands and its neighboring countries maintain their lower fracture risk after migration, or whether these populations approach the fracture incidence of the high-risk host population. Longitudinal changes in vitamin D levels during aging in The Netherlands (N.M. van Schoor): Purpose: Longitudinal changes in serum 25-OHD levels during aging have not been studied extensively. When studying the longitudinal change, it is highly important to adequately adjust for seasonal variation. The objectives are: (1) To examine longitudinal changes in serum 25-OHD levels during aging; and (2) To describe the seasonal variation using a cosine function. Methods: Data of the Longitudinal Aging Study Amsterdam (LASA) were used, an ongoing cohort study. Two different cohorts were included: (1) younger cohort: aged 55–65 years at baseline, n = 738, follow-up of six years; (2) older cohort: aged 65–88 years at baseline, n = 1320, follow-up of thirteen years. Results: In the younger cohort, a longitudinal increase in mean serum 25-OHD levels of 4 nmol/L in six years was observed; in the older cohort, a longitudinal decrease of 4 nmol/L in thirteen years was observed. The seasonal variation was ±12 nmol/L in the younger cohort, and ±7 nmol/L in the older cohort. Conclusions: Long term serum 25-hydroxyvitamin D levels remained fairly stable during ageing with slightly increasing levels in persons aged 55–65 years, and slightly decreasing levels in persons aged 65–88 years. On average, the seasonal variation was larger than the longitudinal change. Associations between hip fractures and psychotropic drugs (M.S. Bakken): Psychotropic drugs are widely used and may cause injurious falls. Aim: To examine associations between the use of antidepressants, anxiolytics and hypnotics among older people and the risk of hip fracture. Methods: A nationwide prospective cohort study of people in Norway born before 1945 (n = 906,422) was conducted. We obtained information on all prescriptions of antidepressants, anxiolytics and hypnotics dispensed in 2004–2010 (the Norwegian Prescription Database) and all primary hip fractures in 2005–2010 (the Norwegian Hip Fracture Registry). We compared the incidence rates of hip fracture during drug exposure and non-exposure by calculating the standardized incidence ratio (SIR). Results: Altogether, 39,938 people (4.4%) experienced a hip fracture. The risk of hip fracture was increased for people exposed to any antidepressant, SIR 1.7 (95% confidence interval 1.7–1.8); tricyclic antidepressants, SIR 1.4 (1.3–1.5) and selective serotonin reuptake inhibitors (SSRIs), SIR 1.8 (1.7–1.8). Preliminary analysis revealed an increased risk of hip fracture among people exposed to anxiolytics, SIR 1.4 (1.4–1.5) and hypnotics, SIR 1.2 (1.1–1.2); the excess risk was highest regarding short-acting benzodiazepine (SAB) anxiolytics, SIR 1.5 (1.4–1.6). Benzodiazepine-like hypnotics (z-hypnotics) were associated with higher excess risk of hip fracture at night, SIR 1.3 (1.2–1.4) than during the day, SIR 1.1 (1.1–1.2). Conclusions: Our nationwide study shows an excess risk of hip fracture among older people using antidepressants, anxiolytics or hypnotics, including the widely prescribed SSRIs, SABs and z-hypnotics (particularly at night). Thus, cautious prescribing is needed.